Tools Plugins Integrations

from crewai_tools import tool

@tool
def add_numbers(num1: int, num2: int) -> int:
    """Adds two numbers and returns the result."""
    return num1 + num2

from crewai import Agent, Task, Crew
from your_tools_module import add_numbers # Assuming your tool is in 'your_tools_module.py'

## Define your agent
calculator_agent = Agent(
    role="Calculator Assistant",
    goal="Perform simple arithmetic calculations",
    backstory="An AI specialized in math.",
    tools=[add_numbers], # Assign the custom tool
    verbose=True
)

## Define a task that uses the tool
calculation_task = Task(
    description="Calculate the sum of 5 and 7.",
    expected_output="The number 12",
    agent=calculator_agent
)

## Create and run the crew
crew = Crew(
    agents=[calculator_agent],
    tasks=[calculation_task],
    verbose=2
)

crew.kickoff()

import requests
from crewai_tools import tool

@tool
def get_weather(location: str) -> str:
    """Fetches the current weather for a given location."""
    api_key = "YOUR_WEATHER_API_KEY" # Replace with your actual API key
    base_url = f"http://api.openweathermap.org/data/2.5/weather?q={location}&appid={api_key}&units=metric"
    try:
        response = requests.get(base_url)
        response.raise_for_status() # Raise an exception for bad status codes
        data = response.json()
        temp = data['main']['temp']
        description = data['weather'][0]['description']
        return f"The weather in {location} is {temp}°C with {description}."
    except requests.exceptions.RequestException as e:
        return f"Error fetching weather for {location}: {e}"

from crewai_tools import ScrapeWebsiteTool

## Initialize the tool
web_scraper = ScrapeWebsiteTool()

## Assign to an agent
researcher_agent = Agent(
    role="Web Researcher",
    goal="Find information on the internet",
    backstory="An expert researcher adept at navigating the web.",
    tools=[web_scraper],
    verbose=True
)

## Define a task
research_task = Task(
    description="Scrape the content of 'https://en.wikipedia.org/wiki/Artificial_intelligence' and summarize the main points about its history.",
    expected_output="A summary of the history of Artificial Intelligence from the Wikipedia page.",
    agent=researcher_agent
)

import os
from crewai import Agent, Task, Crew
from main import YourAgentClass # Example: if your agent class is in 'main.py'

## Set your OpenAI API key as an environment variable
os.environ["OPENAI_API_KEY"] = "YOUR_OPENAI_API_KEY"

## Example of an agent that might be powered by OpenAI models
data_analyst_agent = Agent(
    role="Data Analyst",
    goal="Analyze financial reports and identify trends",
    backstory="An experienced data analyst with a keen eye for detail.",
    verbose=True,
    # By default, CrewAI uses OpenAI models if no LLM is specified.
    # You can explicitly specify an LLM from LangChain if needed.
)

from langchain_community.llms import HuggingFaceHub
from crewai import Agent, Task, Crew
import os

## Set your Hugging Face API token as an environment variable
os.environ["HUGGINGFACEHUB_API_TOKEN"] = "YOUR_HUGGINGFACE_HUB_API_TOKEN"

## Initialize an LLM from Hugging Face
hf_llm = HuggingFaceHub(repo_id="google/flan-t5-large", model_kwargs={"temperature": 0.7, "max_length": 512})

## Define an agent using the Hugging Face LLM
hugging_agent = Agent(
    role="Text Summarizer",
    goal="Summarize provided text accurately",
    backstory="A helpful AI assistant focused on condensing information.",
    llm=hf_llm, # Specify the Hugging Face LLM
    verbose=True
)

## Define a task
summarization_task = Task(
    description="Summarize the following article: [Paste Article Content Here]",
    expected_output="A concise summary of the article.",
    agent=hugging_agent
)

## Create and run the crew
crew = Crew(
    agents=[hugging_agent],
    tasks=[summarization_task],
    verbose=2
)

crew.kickoff()

from crewai_tools import BaseTool, tool, LoadFileTool, CSVSampleTool
from crewai import Agent, Task, Crew

## Example using LoadFileTool
load_file_tool = LoadFileTool()

reader_agent = Agent(
    role="Document Reader",
    goal="Read and process documents",
    backstory="An AI that specializes in extracting information from files.",
    tools=[load_file_tool],
    verbose=True
)

read_task = Task(
    description="Read the content of 'my_document.txt' and extract the key dates mentioned.",
    expected_output="A list of key dates found in the document.",
    agent=reader_agent
)

## This is a conceptual example. Actual implementation requires
## setting up the vector DB and embedding model.

from crewai_tools import tool
## Assume you have a vector_db_client configured for FAISS or Pinecone
## from your_vector_db_setup import vector_db_client

## Placeholder for a tool that searches a vector database
@tool
def semantic_search_vector_db(query: str) -> str:
    """Performs a semantic search in the knowledge base and returns relevant snippets."""
    # Placeholder for actual vector database search logic
    # For example, using Pinecone client:
    # results = vector_db_client.query(vector=query_embedding, top_k=3)
    # return "\n".join([match['metadata']['text'] for match in results['matches']])
    return f"Searching for '{query}' in the knowledge base..."

## Assign this tool to an agent that needs to access a knowledge base
knowledge_agent = Agent(
    role="Knowledge Navigator",
    goal="Find and synthesize information from the company's knowledge base.",
    backstory="An AI with access to a vast repository of internal documents.",
    tools=[semantic_search_vector_db],
    verbose=True
)

## Define a task that requires knowledge retrieval
knowledge_task = Task(
    description="Find information about our latest marketing campaign strategy using the knowledge base.",
    expected_output="Key strategies and findings related to the latest marketing campaign.",
    agent=knowledge_agent
)